Pneumonic and Nonpneumonic Exacerbations of COPD

1134

CHEST Original ResearchCOPD

Original Research

COPD is a growing cause of disease worldwide 1 : according to the Epidemiologic Study of COPD

in Spain (EPI-SCAN) study, the prevalence of COPD in Spain is estimated at 10.2%. 2 Acute exacerbation of COPD (AECOPD) is an event that occurs in the natural course of COPD and requires hospitalization. 3 AECOPD accelerates the decline in lung function 4 and increases risk of death. 5 According to the criteria of Anthonisen et al, 6 AECOPD is defi ned clinically by three symptoms: increased shortness of breath, increased sputum volume, and purulence. Studies have

shown that patients with frequent AECOPD compose a specifi c COPD phenotype. 7

Community-acquired pneumonia (CAP) is an infec-tious disease that is highly prevalent in patients with COPD 8,9 and is probably linked to abnormal host defense mechanisms. 10,11 Elsewhere, we have shown that hospitalized patients with CAP with COPD (CAP 1 COPD) have a different and specifi c early infl amma-tory response compared with patients without COPD. 12

Clinically, in patients with COPD, CAP is not cur-rently considered an exacerbation, even though several

Background: Community-acquired pneumonia (CAP) is a frequent event in patients with COPD, although it is not currently considered an acute exacerbation of COPD (AECOPD). To our knowl-edge, no studies have compared the infl ammatory response of patients with COPD who develop CAP or AECOPD. The aim of our study was to compare clinical and evolutive manifestations and biologic signaling of AECOPD and CAP 1 COPD. Methods: Prospective data were collected from 249 consecutively hospitalized patients with COPD. Comparative analyses were performed in patients with AECOPD (n 5 133) and patients with CAP 1 COPD (n 5 116). Measures of clinical characteristics, blood biomarkers, and evolution were recorded on admission, after 3 and 30 days, and in a follow-up period of 30 days, 90 days, and 1 year. Results: Patients with CAP 1 COPD had higher FEV 1 compared with patients with COPD without pneumonia. In-hospital and long-term outcomes (1 year) were similar for both populations. How-ever, patients with AECOPD had more readmissions, and patients with CAP had more prior epi-sodes of pneumonia. At day 1 and day 3, patients with CAP 1 COPD had signifi cantly ( P , .001) higher serum levels of C-reactive protein (CRP), procalcitonin, tumor necrosis factor- a , and IL-6. Repetition of the analyses after stratifying patients based on severity of disease, current inhaled pharmacotherapy, and noninfectious AECOPD cause confi rmed higher levels of the same bio-markers in patients with CAP 1 COPD. Chills, pleuritic pain, sputum purulence, and CRP levels at day 1 were independent clinical predictors of CAP 1 COPD. Conclusions: Our study confi rms that two different clinical and infl ammatory profi les exist in hos-pitalized patients with COPD in response to CAP (stronger response) and AECOPD, although with similar short-term and long-term outcomes. CHEST 2013; 144(4):1134–1142

Abbreviations: AECOPD 5 acute exacerbation of COPD; AUC 5 area under the curve; CAP 5 community-acquired pneumonia; CRP 5 C-reactive protein; GOLD 5 Global Initiative for Chronic Obstructive Lung Disease; ICS 5 inhaled corti-costeroid; LOS 5 length of hospital stay; NIMV 5 noninvasive mechanical ventilation; PCT 5 procalcitonin; TNF- a 5 tumor necrosis factor a

Pneumonic and Nonpneumonic Exacerbations of COPD Infl ammatory Response and Clinical Characteristics

Arturo Huerta , MD ; Ernesto Crisafulli , MD , PhD , FCCP ; Rosario Menéndez , MD , PhD ; Raquel Martínez , MD ; Néstor Soler , MD , PhD ; Mónica Guerrero , MD ; Beatriz Montull , MD ; and Antoni Torres , MD , PhD , FCCP

Downloaded From: http://journal.publications.chestnet.org/ by a University College London User on 06/24/2014

journal.publications.chestnet.org CHEST / 144 / 4 / OCTOBER 2013 1135

microorganisms such as C pneumoniae ( � 1:32), C burnetii ( � 1:80), and M pneumoniae (any positive titer).

Clinical Evaluations and Determination of Blood Biomarkers

Data on demographic variables were recorded on admission. Symptoms and signs of an acute event together with physiologic and laboratory data were collected at onset. Length of stay (LOS), admission to the ICU, or noninvasive mechanical ventilation (NIMV), and frequency of patients in whom severe sepsis with or without septic shock was identifi ed were also recorded. Mortality and rehospitalization were evaluated during follow-up (30 and 90 days and 1 year). Measurements of C-reactive protein (CRP), procal-citonin (PCT), tumor necrosis factor- a (TNF- a ), and cytokines (IL-1, IL-6, IL-8, and IL-10) were performed on admission and at 3 and 30 days thereafter.

Statistical Analysis

Analysis of variables was performed using a statistical software package (SPSS 17; IBM). Results were expressed as mean � SD or as median (25th-75th percentiles) for continuous variables and as frequency (percentage) for categorical variables.

A prior test for normal data distribution was performed (Kolmogorov-Smirnov test). Differences in continuous variables were analyzed using an independent two-tailed t test for unpaired analysis; otherwise, the nonparametric Mann-Whitney U or Kruskal-Wallis tests were used. Categorical variables were studied using the x 2 test or Fisher exact test when necessary.

To characterize early predictors of a specifi c group of patients, a multivariate analysis of signifi cant clinical and infl ammation var-iables was performed using a stepwise logistic regression model with AECOPD/CAP 1 COPD as a binary dependent variable; continuous variables were categorized by quartiles. A Hosmer-Lemeshow goodness-of-fi t test was also calculated. Finally, the area under a receiver operating characteristic curve (AUC) measured the diagnostic discrimination property of signifi cant predicting biomarkers. For all analyses, P , .05 was considered to be statis-tically signifi cant.

patients present repeated episodes during the course of the disease. 12 Moreover, in clinical practice, it is not always easy to determine whether patients with COPD have pneumonia or a bronchial infection. Mea-surement of increasing plasma levels of infl ammatory biomarkers from the stable condition can be useful to identify an exacerbation and predict mortality 13 ; in addition, no studies have compared the infl amma-tory response of patients with COPD who develop CAP 1 COPD and those who develop AECOPD. The aim of our study was to compare the clinical charac-teristics, systemic infl ammatory profi le, and short-term and long-term evolution between the two groups: CAP 1 COPD and AECOPD.

Materials and Methods

Study Cohort and Defi nitions

We performed a prospective study of consecutive adult patients with COPD hospitalized at two tertiary university hospitals in Spain (Hospital La Fe, Valencia and Hospital Clinic, Barcelona). The study was approved by the ethics committees (Project CEIC 2003/0048 and CEIC 2004/1855 for Hospital La Fe and Hospital Clinic, respectively) in accordance with the Declaration of Helsinki. Written informed consent was obtained from patients. Figure 1 shows the study fl ow diagram. A complete description of COPD, CAP, and AECOPD and sepsis and septic shock defi nitions are provided in e-Appendix 1 .

Microbiologic Evaluation

We attempted to obtain sputum of good quality for bacterial and fungal culture in patients with AECOPD. The microbiologic samples of patients with CAP included sputum samples, two blood cultures, and detection of urinary antigens for Streptococcus pneu-moniae and Legionella pneumophila . Diagnosis of the following microorganisms was also carried out by pairing sera on admis-sion and 3 and 6 weeks after admission: (1) atypical microorganisms (including L pneumophila , Chlamydophila pneumoniae , Chlamydia psittaci , Mycoplasma pneumoniae , and Coxiella burnetii ), and (2) respiratory viruses (infl uenza virus [A and B], parainfl uenza virus, 1-3 syncytial respiratory virus, and adenovirus). High titers of IgM in the acute phase were accepted for diagnosis of atypical

Manuscript received March 6, 2013; revision accepted June 2, 2013 . Affi liations: From the Pneumology Department (Drs Huerta, Soler, and Guerrero and Prof Torres), Clinic Institute of Thorax, Hospital Clinic of Barcelona, Institut d’Investigacions Biomèdiques August Pi i Sunyer, University of Barcelona, Barcelona, CIBERES 06/06/0028, Spain; the Department of Medical and Surgical Sci-ences (Dr Crisafulli), University of Modena and Reggio Emilia, Modena, Italy; and the Pneumology Department (Drs Menéndez, Martinez, and Montull), Hospital Universitario y Politécnico La Fe, CIBERES, Valencia, Spain . Funding/Support: This manuscript was supported by a grant from La Marató TV3 . Correspondence to: Antoni Torres, MD, PhD, FCCP, Pneumology Department, Clinic Institute of Thorax (ICT), Hospital Clinic, Villarroel 170, 08036 Barcelona, Spain; e-mail: [email protected] © 2013 American College of Chest Physicians. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details. DOI: 10.1378/chest.13-0488

Figure 1. Study fl ow diagram. AECOPD 5 acute exacerbation of COPD; CAP 5 community-acquired pneumonia .


http://journal.publications.chestnet.org

mailto:[email protected]

1136 Original Research

the potential effect of home-based pharmacotherapy (for short-acting bronchodilators, data not shown; for inhaled corticosteroids [ICSs], see Fig 3 ) confi rmed higher levels of CRP, PCT, TNF- a , and IL-6 in patients with CAP 1 COPD, especially on admission (day 1 and day 3).

The logistic regression analysis ( Fig 4 ) performed with the specifi c diagnosis (AECOPD or CAP 1 COPD) as a dependent variable and adjusted for descriptive confounders (GOLD [Global Initiative for Chronic

Results

Data on 249 patients with COPD admitted to two tertiary university hospitals were collected. Of these, 133 patients (53%) were considered to have AECOPD, and 116 patients (47%) had CAP associated with COPD. Severity of COPD according to lung function (mean FEV 1 , 47% � 16% predicted), number of patients requiring long-term oxygen therapy (n 5 27, 20.3%), and use of short-acting bronchodilators were greater in patients with AECOPD. Table 1 shows the main characteristics.

Clinical features of patients with COPD at hospi-talization and during follow-up ( Tables 2, 3 ) were as follows: Patients with CAP 1 COPD had higher prev-alence of symptoms/signs (fever, chills, sputum puru-lence, pleuritic pain) and severe sepsis, whereas patients requiring NIMV and patients readmitted in a period of 30 days were more prevalent in the AECOPD group. Other characteristics of clinical severity (Pa o 2 /F io 2 , LOS, admission to ICU, and septic shock) were sim-ilar in both groups. Furthermore, the mortality calcu-lated on hospitalization and at 30 and 90 days and at 1 year was also similar. The prevalence of recurrent acute events showed a higher proportion of patients with a prior episode of pneumonia in patients with CAP 1 COPD ( P , .001) and more patients with more than two exacerbations per year in the AECOPD group ( P , .01 for both variables in comparison with CAP 1 COPD) ( Fig 2 ).

Rate of microbiologic diagnosis ( Table 2 ) was con-fi rmed in 44% and 47% of patients with AECOPD and patients with CAP 1 COPD, respectively. In the etiology comparison, Haemophilus infl uenzae was more prevalent in AECOPD (19% vs 5% of CAP 1 COPD, P 5 .04), whereas S pneumoniae was more prevalent in CAP 1 COPD (43% vs 10% of AECOPD, P , .001). Other microbial causes were similar between groups.

On admission and 3 days afterward, patients with CAP 1 COPD had signifi cantly higher biologic sig-naling in comparison with patients with AECOPD ( P , .001 for CRP, PCT, TNF- a , IL-6; P , .05 for IL-1, IL-8). On day 30, levels of IL-6 were higher ( P , .05) in patients with CAP 1 COPD. The remain ing IL-10 level was similar in both groups ( Table 4 ).

A new analysis performed with the exclusion of patients without a confi rmed infectious cause from the AECOPD group (AECOPD, n 5 59; CAP 1 COPD, n 5 116) (e-Table 1) showed similarly higher levels of biomarkers in the CAP 1 COPD group ( P , .001 for CRP, TNF- a , IL-6; P , .05 for PCT).

Analysis of signaling according to the severity of air-fl ow obstruction confi rmed higher levels of CRP, PCT, TNF- a , and IL-6 ( P , .05) in patients with CAP 1 COPD in comparison with AECOPD (e-Table 2). Repetition of the analyses of biomarkers based on

Table 1— General Characteristic of Study Cohorts

Characteristic Patients With

AECOPDPatients With CAP 1 COPD P Value

No. 133 116Age, y 69.4 � 9.8 71.9 � 10.0 .06Sex, male (female) 124 (9) 111 (5) .58Smoking habit Nonsmokers/current/

former, %3.8/28.0/68.2 7.8/30.2/62.1 .33

Alcohol habit No alcohol/current/

former , %80.2/14.5/5.3 81.0/10.3/8.6 .46

FEV 1 , L 1.30 � 0.6 1.44 � 0.5 .04 FEV 1 % predicted 47.8 � 16.0 53.2 � 18.2 .01FEV 1 /FVC, % 46.9 � 12.2 51.7 � 13.6 , .01COPD-GOLD

stage I/II/III/IV, %2/37/37/24 10/46/34/10 , .01

LTOT 27 (20.3) 7 (6.0) , .001Chronic comorbidities Congestive heart failure 32 (24.0) 34 (29.3) .56 Chronic renal failure 2 (1.5) 6 (5.2) .11 Diabetes 20 (15.0) 26 (22.4) .28 Noncirrhotic liver

disease4 (3.0) 4 (3.4) .84

Neurologic disease 10 (7.5) 16 (13.8) .14 Nonactive cancer 12 (9.0) 5 (4.3) .20Dyspnea

grade a 0/1/2/3/4, %1/7/32/54/6 3/10/39/41/7 .34

Home pharmacotherapy SABA 90 (67.6) 54 (46.6) , .001 Salbutamol/terbutaline 86/4 50/4 .45 LABA 81 (60.9) 45 (38.8) , .001 Formoterol/salmeterol 26/55 11/34 .41 Anticholinergics 91 (68.4) 53 (45.7) , .001 Ipratropium/tiotropium 37/54 28/25 .15 ICS 75 (56.3) 62 (53.4) .64 Budesonide/fl uticasone 34/41 27/35 .83 OCS 12 (9.0) 13 (11.2) .56 Prednisone/

methylprednisone/defl azacort

6/2/4 5/5/3 .40

Vaccination Antipneumococcal 27 (20.3) 22 (19.0) .79 Infl uenza 82 (61.6) 60 (51.7) .11

Values are reported as mean � SD or as frequency (%). AECOPD 5 acute exacerbation of COPD; CAP 5 community-acquired pneumonia; GOLD 5 Global Initiative for Chronic Obstructive Lung Disease; ICS 5 inhaled corticosteroid; LABA 5 long-acting b 2 -agonist; LTOT 5 long-term oxygen therapy; OCS 5 oral corticosteroid; SABA 5 short-acting b 2 -agonist. a Assessed by Medical Research Council scale.



0.71; 95% CI, 0.65-0.78; P , .001) with a best cutoff of 12.9 mg/dL (0.623 and 0.628 in the sensitivity and specifi city evaluation) ( Fig 5 ).

Discussion

The main fi ndings of this study are as follows: (1) The clinical characteristics of both patients with AECOPD and patients with CAP 1 COPD on admis-sion were different. Patients with AECOPD had greater severity of disease than patients with CAP 1 COPD. (2) Bacterial microbiology was different, with a predom-inance of H infl uenzae in the AECOPD group and S pneumoniae in the CAP 1 COPD group. (3) Despite different levels of severity of COPD, short-term and

Obstructive Lung Disease] classifi cation, current domi-ciliary use of short-acting bronchodilators, and infec-tious cause) showed that clinical symptoms/signs (chills, pleuritic pain, and purulence sputum production) were clinical predictors for performing early discrim-ination of patients with CAP 1 COPD from patients with AECOPD. Moreover, CRP on day 1 (levels cate-gorized by quartiles with � 3.5 mg/dL, fi rst quartiles used as references; 3.6-11.9 mg/dL, second quartiles; 12-20.4 mg/dL, third quartiles; � 20.5 mg/dL, fourth quartiles) showed an increasing OR in the prediction of CAP (OR, 4.02; 95% CI, 1.8-9.9; OR, 5.75; 95% CI, 3.2-12.7; and OR, 10.6; 95% CI, 4-25.4 in the second, third, and fourth quartiles, respectively; P , .001 for all quartiles). The accuracy analysis of CRP on day 1 showed a signifi cant predictive discrimination (AUC,

Table 2— Clinical, Physiologic, Laboratory, and Microbiologic Features Measured at Onset of Exacerbation

Feature Patients With AECOPD Patients With CAP 1 COPD P Value

Fever 51 (38.3) 81 (69.8) , .001Chills 38 (28.5) 73 (62.9) , .001Cough 107 (80.4) 94 (81) .90Sputum categories: absence/mucous/purulence/rusty, % 31/53/14/2 24/39/30/7 , .001Pleuritic pain 17 (12.7) 42 (36.2) , .001Dyspnea 131 (98.4) 95 (81.9) , .001Respiratory rate, breaths/min 30 (27-35) 27 (24-31 ) .31Heart rate, beats/min 100 (85-112) 100 (85-110.5) .60Systolic BP, mm Hg 126 (114-142) 139 (125-148) .26Diastolic BP, mm Hg 75 (58-79) 76 (64.5-86) .01Leukocytes counts, 10 9 cell/L 12,150 (10,700-17,675) 15,100 (11,200-19,500) , .001Hematocrit, % 42.5 (37-44) 4.4 (41.4-46.2) , .001Glycemia, mg/dL 125.5 (106-138.2) 133 (107.5-155.7) .25Creatinine, mg/dL 1.1 (0.95-1.6) 1 (0.95-1.4) .02Sodium 137 (132.7-139) 135 (133-138) , .001Potassium 4.5 (4.1-4.8) 4.2 (3.9-4.6) .04pH 7.40 (7.37-7.43) 7.43 (7.39-7.46) , .001Pa o 2 /F io 2 266 (223.5-318.5) 279.5 (244.5-300) .34Pa co 2 , mm Hg 47 (41.1-56.9) 40.6 (34.3-50) , .001HCO 3 , mmol/L 28.9 (25.8-31.9) 26.8 (24-29.7) .001BE, mmol/L 3.3 (1.35-6) 2.5 (0.45-4) .01Confi rmed infective pathogen n 5 59 (44%) n 5 55 (47%) .70 Streptococcus pneumoniae 6 (10) 24 (43) , .001 Staphylococcus aureus a 5 (8) 4 (7) .93 Haemophilus infl uenzae 11 (19) 3 (5) .04 Haemophilus parainfl uenzae 4 (7) 1 (2) .24 Moraxella catarrhalis 3 (5) 1 (2) .61 Legionella pneumophila 1 (2) 2 (4) .46Other atypical pathogens b 3 (5) 4 (7) .54 Acinetobacter species 3 (5) 1 (2) .39 Respiratory virus 4 (7) 0 (0) .06 Pseudomonas aeruginosa 10 (17) 8 (15) .80 Enterobacteriaceae c 6 (10) 5 (9) .97 Aspergillus species 1 (2) 0 (0) .35Others 2 (3) 2 (4) .86

Values are reported as mean � SD, median (25th-75th percentiles), or frequency (%); for microbiologic data values are reported as frequency (%) in relation to the number of patients with etiologic diagnosis in each group. BE 5 base excess; HCO 3 5 serum bicarbonate. See Table 1 legend for expansion of other abbreviations. a Including methicillin-sensitive and methicillin-resistant Staphylococcus aureus . b Including Mycoplasma pneumoniae , Coxiella burnetii , and Chlamydia . c Including Escherichia coli , Klebsiella oxytoca , and Proteus mirabilis.




logic response may be used as diagnostic marker in some cases in which the infi ltrates on the chest radio-graph are in doubt.

Clinical and Microbiologic Differences

The functional characteristics of the two cohorts stud-ied were dissimilar, as the AECOPD cohort included patients with more severe disease. We also observed differences, with more pleuritic pain, fever, chills, and leukocytosis in the CAP 1 COPD group. However, the severity of the acute disease was similar in terms of oxygenation and ICU admission. A prior study with a smaller number of patients, published . 10 years ago and comparing similar cohorts, found the same FEV 1 for both populations. 14 We do not have a clear explanation for this discrepancy, but it may be that patients hospitalized with AECOPD . 10 years ago had less severe disease than now (in terms of FEV 1 ).

Regarding microbiologic cause, our study con-fi rms the different and specifi c prevalence of path-ogens in patients with AECOPD and patients with CAP 1 COPD 15 : More H infl uenzae was associated with AECOPD, 16 whereas pneumococcal causes were more frequent in cases of pneumonia. P aeruginosa

long-term outcomes, including mortality, were sim-ilar. The only exception was the readmission rate in patients with AECOPD compared with patients with CAP 1 COPD (24% and 12%, respectively). (4) We showed that bacterial pneumonia infection produces more powerful biologic signaling, mediated by sev-eral biomarkers (CRP, PCT, TNF- a , and IL-6): This response is not infl uenced by severity of COPD or the use of current pharmacotherapy, particularly inhaled corticosteroids, or presence of infectious cause. (5) In our multivariate model, we showed that, in addi-tion to initial symptoms, such as chills, pleuritic pain, and purulence of sputum, higher levels of CRP on admission is a strong predictor for discriminating CAP 1 COPD from AECOPD; this characteristic bio-

Table 3— Clinical Outcomes Evaluated During Hospital Course and During Follow-up

VariablesPatients With


Mental status alteration 11 (8.3) 11 (9.5) .82LOS, d 9.5 � 8.2 10.1 � 6.5 .53Require NIMV 8 (6.0) 1 (0.9) .03ICU admission 5 (3.8) 3 (2.6) .72Severe sepsis … 18 (15.5) …Septic shock … 4 (3.4) …Mortality During hospitalization 4 (3.0) 4 (3.4) .99 On day 30 5 (3.8) 4 (3.4) .97 On day 90 7 (5.3) 4 (3.4) .55 At 1 y 20 (15.0) 14 (12.1) .58Rehospitalization for a

new episode In a period of 30 d 32 (24.1) 14 (12.1) .02 In a period of 90 d 33 (24.8) 25 (21.6) .55 In a period of 1 y 53 (39.8) 51 (44.0) .52

Values are reported as mean � SD or as frequency. LOS 5 length of hospital stay; NIMV 5 noninvasive and invasive mechanical ventilation. See Table 1 legend for expansion of other abbreviations.

Figure 2. Prevalence of patients according to type of exacerba-tion. Open bars 5 patients with AECOPD; fi lled bars 5 patients with CAP 1 COPD. § P , .01; §§ P , .001. See Figure 1 legend for expansion of abbreviations.

Table 4— Panel of Infl ammatory Response

BiomarkerPatients With


CRP, mg/dL Day 1 6.9 (1.3-17.0) 16.8 (8.4-25) , .001 Day 3 1.3 (0.3-3.6) 4 (1.8-9.4) , .001 Day 30 0.5 (0.1-2) 0.9 (0.3-2.5) .11PCT, ng/mL Day 1 0.09 (0.09-0.40) 0.30 (0.09-1.93) , .001 Day 3 0.09 (0.09-0.35) 0.13 (0.09-0.43) .02 Day 30 0.09 (0.09-0.09) 0.09 (0.09-0.09) .97TNF- a , pg/mL Day 1 12 (7-22) 17 (13-28.2) , .001 Day 3 12.2 (7-22) 15.5 (11-24) , .01 Day 30 14 (9-19.5) 15 (10.7-26) .35IL-1, pg/mL Day 1 3.4 (0-10) 4 (3-16) .03 Day 3 3 (0-8) 5 (3-19) , .001 Day 30 3 (0-7.5) 4 (0-8) .14IL-6, pg/mL Day 1 22 (0-134) 84.5 (24-230) , .001 Day 3 5.5 (0-32) 21 (11-42.7) , .001 Day 30 13 (5-33) 19.5 (9.7-34) .02IL-8, pg/mL Day 1 3 (0-17) 11 (4-20.7) , .001 Day 3 0 (0-6.5) 2 (0-9.2) , .01 Day 30 8 (0-22.5) 9.50 (2-22.2) .13IL-10, pg/mL Day 1 0 (0-9) 0 (0-9.2) .94 Day 3 9 (0-18) 9.5 (4.7-17.2) .31 Day 30 0 (0-8) 0 (0-8.2) .74

Values are reported as median (25th-75th percentiles). CRP 5 C reactive protein; PCT 5 procalcitonin; TNF- a 5 tumor necrosis factor- a . See Table 1 legend for expansion of other abbreviations.



exacerbation on patients survival, showed that mortality 6 months after discharge from the hospital among patients readmitted one, two, or more times was 27%, 31%, and 36%, respectively, whereas those requiring no readmission had a mortality rate of 21%. 5 Method-ologic differences in data collection might explain our results on lower mortality at 1 year of follow-up but undoubtedly require further studies.

The only difference observed was a 24% readmis-sion rate for patients with AECOPD compared with 12% for patients with CAP 1 COPD. This is probably explained by a higher baseline level of COPD severity in patients with AECOPD. In addition, this is a major issue in terms of health-care cost. Studies identifying risk factors for short-term readmission are needed in both populations.

was similarly distributed. These differences may be explained by the different periods of COPD evolution.

Outcomes

Although this study is not powered for mortality, we found that mortality rates during hospitalization and in short-term and long-term follow-up were sim-ilar between groups, despite higher COPD functional severity of AECOPD compared with CAP 1 COPD, but lower than expected for both populations. For example, at 1 year, we observed 15% and 12% mortal-ities when comparing AECOPD to CAP 1 COPD, respectively. Studies done in pure populations of patients with AECOPD, included from the stable phase of the disease with the aim of assessing the effect of

Figure 3. Panel of signifi cant infl ammatory biomarkers evaluated in all patients and according to domi-ciliary inhaled corticosteroid use. Open bars 5 patients with AECOPD; fi lled bars 5 patients with CAP 1 COPD. The horizontal bar and box length represent the median and the interquartile range, respectively. Circles and asterisks indicate outliers (1.5 and 3 times the interquartile range, respectively). § P , .05; §§ P , .001. CRP 5 C-reactive protein; PCT 5 procalcitonin; TNF- a 5 tumor necrosis factor- a . See Figure 1 legend for expansion of other abbreviations.




suggest the innate immunologic response of different infl ammatory COPD phenotypes. 10,21,22

Prediction of Pneumonia

Currently, an exacerbation of COPD is diagnosed in clinical settings and confi rmed or excluded by means of chest radiograph. However, symptoms of presenta-tion such as fever and productive cough are common in both AECOPD and CAP, and the incidence of pneumonia in patients already cataloged with the diagnosis of AECOPD may be higher. 14 Moreover, a chest radiograph may not be suffi ciently sensitive to identify a new pulmonary infi ltrate. In those cases, physicians often need to request a CT scan, which is not often practical, and prescribe unneeded or non-specifi c antibiotics.

Our logistic regression model confi rms that the pres-ence of chills, pleuritic pain, and sputum purulence are symptoms predictive of CAP. Based on the ratio-nale of reducing antibiotic prescriptions, the study by Bafadhel et al 23 found that procalcitonin and CRP can independently distinguish pneumonia from exacerba-tions of asthma or COPD; however, caution has been voiced regarding procalcitonin as diagnostic marker

Characterization of Specifi c Infl ammatory Profi les

In this study, we demonstrated that patients with AECOPD have a different pattern of infl ammatory response to exacerbation, with lower levels of CRP, PCT, TNF- a , and IL-6, especially in the fi rst 3 days after admission. Specifi cally in terms of a long-term infl ammatory response, we note with interest that levels of IL-6 were also higher in patients with CAP 1 COPD at day 30. Neither a higher initial or late inflam-matory response in patients with CAP 1 COPD was associated with a higher mortality compared with AECOPD.

We performed three subanalyses of the infl amma-tory response, one categorizing the patients accord-ing to COPD severity, one excluding patients with AECOPD with potential noninfective cause, and one taking into account prior administration of ICS. We found similar results when analyzing the overall pop-ulation, with a higher infl ammatory response at days 1 and 3 in patients with CAP 1 COPD and a higher resid-ual infl ammation at day 30 in this population.

Regarding the hypothetical baseline bias of severity of COPD (based on local infl ammation), 17 use of cur-rent pharmacotherapy (especially ICS), 18 and presence of an infective etiology in nonpneumonic exacerba-tions, 19,20 we found no modulating effect on the mea-surement of early infl ammatory response (e-Table 2, Fig 3, and e-Table 1, respectively); these results may

Figure 4. Stepwise logistic regression analysis model performed with AECOPD/CAP 1 COPD as a dependent variable. §Categories defi ned by quartiles; * P , .05; ** P , .001. The fi lled squares and solid lines represent OR values and 95% CIs. Model analysis adjusted for COPD-GOLD (Global Initiative for Chronic Obstruc-tive Lung Disease) staging, current domiciliary use of short-acting b -agonist, long-acting b -agonist, anticholinergics, inhaled cortico-steroids, and infectious cause. Hosmer-Lemeshow goodness-of-fi t test: x 2 , 12.6; P 5 .245. See Figure 1 and 3 legends for expansion of abbreviations.

Figure 5. Receiver operating curve of CRP at day 1, performed with patients with CAP 1 COPD as test variable. Gray line repre-sents a diagonal of reference. AUC 5 area under the curve. See Figure 3 legend for expansion of other abbreviation.



in bacterial infection of AECOPD. 24 Signifi cantly, in our study, high levels of CRP on admission to the hos-pital may help in the diagnosis of one acute event, thereby discriminating patients with CAP 1 COPD from patients with AECOPD ( Figs 4, 5 ). For this rea-son, we think that quick and easy detection of CRP may be useful in some cases (eg, doubtful infi ltrates, condensation in patients with emphysema), with the possibility of screening for a new pulmonary infi ltrate.

Strengths and Limitations

Although a previous study 14 has compared the clin-ical manifestations and disease course of AECOPD and CAP 1 COPD, to our knowledge, our study cohort is the fi rst in which two specifi c infl ammatory patterns have been described. Furthermore, diagnosis of COPD performed at least 6 months prior to hospital admission and in a stable phase of the disease was confi rmed by spirometry; this functional approach correctly diag-noses COPD. Another strength is that we were able to follow the patients up to 1 year after discharge.

One limitation is that we did not know the infl am-mation levels prior to the acute event in patients with either AECOPD or CAP 1 COPD. Although this may have little clinical relevance, recent publications sug-gest that some patients with COPD have a chronic profi le of high levels of cytokines. 25 Another limitation of our study is that we did not perform follow-up serial pulmonary function tests and we do not know whether pulmonary functional impairment is similar or differ-ent in the two populations.

In conclusion, our study shows that AECOPD and CAP 1 COPD are different acute events in terms of baseline severity, microbiology, and infl ammatory pat-tern. Although short-term and long-term mortality is similar, we think that patients with CAP 1 COPD should not be categorized as AECOPD, as has been done in several studies.

Acknowledgments Author contributions: The authors take responsibility and vouch for the completeness and accuracy of the data and analyses. Dr Huerta and Prof Torres are the guarantors of the entire manuscript. Dr Huerta: contributed to the study concept and design, coordi-nation of the acquisition of data, data analysis, interpretation of the data, and writing the article. Dr Crisafulli: contributed to the study concept and design, data analysis, interpretation of the data, and writing the article. Dr Menéndez: contributed to the study concept and design, data analysis, interpretation of the data, and critical revision of the manuscript. Dr Martínez: contributed to the study concept and design, data analysis, interpretation of the data, data analysis, and writing the article. Dr Soler: contributed to the study concept and design, interpreta-tion of the data, and critical revision of the manuscript. Dr Guerrero: contributed to the study concept and design, coor-dination of the acquisition of data, interpretation of the data, and writing the article.

Dr Montull: contributed to the study concept and design, coordi-nation of the acquisition of data, interpretation of the data and writing the article. Prof Torres: contributed to the study concept and design, interpre-tation of the data, and writing and critically revising the manuscript. Financial/nonfi nancial disclosures: The authors have reported to CHEST that no potential confl icts of interest exist with any companies/organizations whose products or services may be dis-cussed in this article. Role of sponsors : The sponsor had no role in the design of the study, the collection of and analysis of the data, or in the prepara-tion of the manuscript. Other contributions: The study was performed at Hospital Clinic, Barcelona, Spain and Hospital Universitario y Politécnico La Fe, Valencia, Spain. Additional information: The e-Appendix and e-Tables can be found in the “Supplemental Materials” area of the online article.

References 1 . Buist AS , McBurnie MA , Vollmer WM , et al ; BOLD Collab-

orative Research Group . International variation in the prev-alence of COPD (the BOLD Study): a population-based prevalence study . Lancet . 2007 ; 370 ( 9589 ): 741 - 750 .

2 . Soriano JB , Miravitlles M , Borderías L , et al . Geographical var-iations in the prevalence of COPD in Spain: relationship to smoking, death rates and other determining factors [in Spanish] . Arch Bronconeumol . 2010 ; 46 ( 10 ): 522 - 530 .

3 . Wedzicha JA , Seemungal TA . COPD exacerbations: defi ning their cause and prevention . Lancet . 2007 ; 370 ( 9589 ): 786 - 796 .

4 . Donaldson GC , Seemungal TA , Bhowmik A , Wedzicha JA . Rela tionship between exacerbation frequency and lung func-tion decline in chronic obstructive pulmonary disease . Thorax . 2002 ; 57 ( 10 ): 847 - 852 .

5 . Soler-Cataluña JJ , Martínez-García MA , Román Sánchez P , Salcedo E , Navarro M , Ochando R . Severe acute exacerba-tions and mortality in patients with chronic obstructive pul-monary disease . Thorax . 2005 ; 60 ( 11 ): 925 - 931 .

6 . Anthonisen NR , Manfreda J , Warren CP , Hershfi eld ES , Harding GK , Nelson NA . Antibiotic therapy in exacerbations of chronic obstructive pulmonary disease . Ann Intern Med . 1987 ; 106 ( 2 ): 196 - 204 .

7 . Hurst JR , Vestbo J , Anzueto A , et al ; Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) Investigators . Susceptibility to exacerbation in chronic obstructive pulmonary disease . N Engl J Med . 2010 ; 363 ( 12 ): 1128 - 1138 .

8 . Soriano JB , Visick GT , Muellerova H , Payvandi N , Hansell AL . Patterns of comorbidities in newly diagnosed COPD and asthma in primary care . Chest . 2005 ; 128 ( 4 ): 2099 - 2107 .

9 . Müllerova H , Chigbo C , Hagan GW , et al . The natural his-tory of community-acquired pneumonia in COPD patients: a population database analysis . Respir Med . 2012 ; 106 ( 8 ): 1124 - 1133 .

10 . Gutierrez P , Closa D , Piñer R , Bulbena O , Menéndez R , Torres A . Macrophage activation in exacerbated COPD with and without community-acquired pneumonia . Eur Respir J . 2010 ; 36 ( 2 ): 285 - 291 .

11 . Taylor AE , Finney-Hayward TK , Quint JK , et al . Defective macrophage phagocytosis of bacteria in COPD . Eur Respir J . 2010 ; 35 ( 5 ): 1039 - 1047 .

12 . Crisafulli E , Menéndez R , Huerta A , et al. Systemic infl amma-tory pattern of patients with community-acquired pneumonia with and without COPD. Chest. 2013 ;143(4):1009-1017.

13 . Celli BR , Locantore N , Yates J , et al ; ECLIPSE Investigators . Infl ammatory biomarkers improve clinical prediction of mor-tality in chronic obstructive pulmonary disease . Am J Respir Crit Care Med . 2012 ; 185 ( 10 ): 1065 - 1072 .




14 . Lieberman D , Lieberman D , Gelfer Y , et al . Pneumonic vs non pneumonic acute exacerbations of COPD . Chest . 2002 ; 122 ( 4 ): 1264 - 1270 .

15 . Li XJ , Li Q , Si LY , Yuan QY . Bacteriological differences between COPD exacerbation and community-acquired pneumonia . Respir Care . 2011 ; 56 ( 11 ): 1818 - 1824 .

16 . Rosell A , Monsó E , Soler N , et al . Microbiologic determinants of exacerbation in chronic obstructive pulmonary disease . Arch Intern Med . 2005 ; 165 ( 8 ): 891 - 897 .

17 . Hogg JC , Chu F , Utokaparch S , et al . The nature of small-airway obstruction in chronic obstructive pulmonary disease . N Engl J Med . 2004 ; 350 ( 26 ): 2645 - 2653 .

18 . Antoniu SA . Effects of inhaled therapy on biomarkers of sys-temic infl ammation in stable chronic obstructive pulmonary disease . Biomarkers . 2010 ; 15 ( 2 ): 97 - 103 .

19 . Menéndez R , Sahuquillo-Arce JM , Reyes S , et al . Cytokine activation patterns and biomarkers are infl uenced by micro-organisms in community-acquired pneumonia . Chest . 2012 ; 141 ( 6 ): 1537 - 1545 .

20 . Sapey E , Stockley RA . COPD exacerbations. 2: aetiology . Thorax . 2006 ; 61 ( 3 ): 250 - 258 .

21 . Hodge S , Hodge G , Scicchitano R , Reynolds PN , Holmes M . Alveolar macrophages from subjects with chronic obstructive pulmonary disease are defi cient in their ability to phagocytose apoptotic airway epithelial cells . Immunol Cell Biol . 2003 ; 81 ( 4 ): 289 - 296 .

22 . Gordon S , Martinez FO . Alternative activation of macro-phages: mechanism and functions . Immunity . 2010 ; 32 ( 5 ): 593 - 604 .

23 . Bafadhel M , Clark TW , Reid C , et al . Procalcitonin and C-reactive protein in hospitalized adult patients with com-munity-acquired pneumonia or exacerbation of asthma or COPD . Chest . 2011 ; 139 ( 6 ): 1410 - 1418 .

24 . Falsey AR , Becker KL , Swinburne AJ , et al . Utility of serum procalcitonin values in patients with acute exacerbations of chronic obstructive pulmonary disease: a cautionary note . Int J Chron Obstruct Pulmon Dis . 2012 ; 7 : 127 - 135 .

25 . Agustí A , Edwards LD , Rennard SI , et al ; Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE) Investigators . Persistent systemic infl ammation is associated with poor clinical outcomes in COPD: a novel phenotype . PLoS ONE . 2012 ; 7 ( 5 ): e37483 .


Documents

Pneumonic and Nonpneumonic Exacerbations of COPD